20:00:47 #startmeeting 20:00:47 Meeting started Tue Dec 23 20:00:47 2014 UTC. The chair is h01ger. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:47 Useful Commands: #action #agreed #help #info #idea #link #topic. 20:00:51 #chair MoC 20:00:51 Current chairs: MoC h01ger 20:00:57 #chair intrigeri 20:00:57 Current chairs: MoC h01ger intrigeri 20:00:58 hmm, I should read the status update first. 20:01:17 intrigeri: do that while i explain the basics of meetbot 20:02:02 MoC: have a look at http://meetbot.debian.net/apparmor/2014/ - pretty boring :) once you look, i will do something... 20:02:16 s/#once you look#once you looked# 20:02:44 h01ger: ok 20:02:51 # topic init - collect agenda 20:02:56 #topic init - collect agenda 20:03:03 #save 20:03:10 look again :) 20:03:27 nice 20:03:49 http://meetbot.debian.net/apparmor/2014/apparmor.2014-12-23-20.00.html has all the #info #agreed stuff 20:03:56 #agreed meetbot rocks :) 20:03:58 #save 20:04:22 http://meetbot.debian.net/apparmor/2014/apparmor.2014-12-23-20.00.* is updated at the end of the meeting end on #save 20:04:58 do i add an agenda topic by doing the topic command? 20:05:17 no, normally meetbot changes channel topic 20:05:23 oh ok 20:06:06 OK, read the report and had a look at the recent wiki changes. 20:06:12 intrigeri: great 20:06:42 general impression: wow, you're getting up to speed! :) 20:06:48 intrigeri: hehehe :) 20:07:00 jmichael_: sbeattie1 sarnold_ kees: could you please set +t so meetbot can change topics or add me to the channel access list, so i can do +t before a meeting and -t after... 20:07:05 so how shoudl we structure this meeting? 20:07:20 collecting an agenda, going through the topics :) 20:07:34 talk about the documentation i wrote and other done stuff? 20:07:40 then todo ? then next meeting? 20:07:44 yup. 20:08:05 status. docs. todo. next meeting. - those 4 ? anything else? 20:08:11 reports 20:08:15 and also define when to send status updates :) 20:08:23 that's it. 20:08:24 reports is status for me, no? 20:08:31 yes 20:08:37 well, no 20:08:48 current status != status report deadlines? 20:09:06 that's what I meant with "reports" :) 20:09:11 shall we start with the deadline for the reports? this might be the easiest topic 20:09:12 so. 1. current status 2. status report deadlines 3. docs 4. todo 5. next meetings. right? 20:09:28 * h01ger is very fine with switching 1+2 :) 20:09:30 ok 20:09:33 more topics? 20:09:49 nothing I can think of right now. 20:10:03 #agreed 1. status report deadlines 2. current status 3. docs 4. todo 5. next meetings. 6. any other business 20:10:07 #save 20:10:20 #topic status report 20:10:23 sorry 20:10:26 #topic status report deadlines 20:10:36 please tell me what's best for you 20:11:21 I need >= 24h to process the report and be useful at a meeting (as you'll see today, I've not had time to think of any of this yet, 20:11:25 how many hours before a meeting do you want the reports or shall i do them every friday? 20:11:26 no time to follow links etc.) 20:11:45 but >48h would be meaningless (there's too much new stuff that can happen in the last 48h), probably 20:11:50 yes 20:11:52 so I would say between 24 and 48h. 20:11:57 and h01ger? 20:12:04 * h01ger is fine with that 20:12:07 ok 20:12:09 me too 20:12:13 but if exceptionally it's 18h or 56h, it's fine too :) 20:12:18 :D 20:12:27 its also fine, rather obvious, if the latest stuff is missing then 20:12:34 might give for more interesting updates during the meeting 20:12:41 yes 20:12:42 also, I'd rather see these reports sent to pkg-apparmor-team. 20:12:47 +yes, exceptions rules the world ;) 20:12:58 * h01ger nods intri and public reports 20:13:34 also, first time we're having a meeting *here*, so to anyone reading: 20:13:55 that's a mentoring/coordination meeting about MoC's OPW project: https://wiki.debian.org/OutreachProgramForWomen/UlrikeU 20:13:58 . 20:14:19 MoC: fine with emailing reports to the list in the future? 20:14:22 ok, fine with me 20:14:30 Cool. next topic, then! 20:14:35 i can also send today's one 20:14:55 MoC: if you want, otherwise no big deal. the future is more important :) 20:15:01 #agreed status reports should (in general) be send >24h in advance to our irc meetings and be send to the pkg-apparmor-team mailinglist. 20:15:01 :D 20:15:12 #save 20:15:14 krkrk 20:15:15 i think re-sending the current/old one is nice 20:15:29 #topic status updates 20:16:18 ulrikes blog is on http://planet.debian.org now :) 20:16:48 ok, so the biggest thing i did this week is work on https://wiki.debian.org/AppArmor/Progress and https://wiki.debian.org/AppArmor/PackageMaintainers 20:18:02 on the blog itself i did simply explain how a profile gets into debian and how to checkout the launchpad profiles and modify them. but i think the doc might need some review. 20:18:09 [not read the blog posts yet] 20:18:28 MoC: these docs are really great! 20:18:44 thanks h01ger 20:18:52 I'll review it later. 20:18:57 great 20:19:13 what i am still having a hard time with is defining the usertags 20:19:22 or at least propose some 20:19:31 #info usertags still unclear 20:19:33 #save 20:19:39 i guess i can take some more time and try to discuss this on the mailing list? 20:19:49 #link https://wiki.debian.org/AppArmor/Progress 20:19:50 iirc, last time we agreed that the usertags would follow once we have the user stories. 20:19:56 yes 20:19:57 #link https://wiki.debian.org/AppArmor/PackageMaintainers 20:20:12 #link https://apparmor.451f.org/2014/12/23/how-does-my-apparmor-profile-get-into-debian/ 20:20:15 so I'm happy to discuss this, once we have the user stories drafted and agreed upon. 20:20:20 (on the list :) 20:20:21 #link https://apparmor.451f.org/2014/12/23/how-to-contribute-to-the-apparmor-upstream-profiles/ 20:20:28 intrigeri: ack 20:20:31 makes sense 20:20:56 #info usertags should follow once we have the user stories 20:20:58 #save 20:21:04 MoC: was time the only blocker wrt. working on the user stories, or is there anything else (and especially anything we can help with)? 20:21:15 intrigeri: good that you ask! 20:21:21 intrigeri: no, time was not the only blocker 20:21:29 (I understand that doing everything you did was anyway needed first, to get an idea of the big picture) 20:21:44 intrigeri: also, i have a hard time imagining more, although i added one story based upon the mysql merge request 20:22:12 * h01ger thinks its ok to just have a few stories now 20:22:25 * intrigeri looking at https://wiki.debian.org/AppArmor/Progress#User_Stories 20:22:32 i started adding some proposals for tags in italic for each story 20:22:37 ... and really confused by the formatting. 20:22:39 but not yet much 20:22:40 hopefully there will be 2-3 more in 3 months, but i dont see 30 stories anytime 20:22:53 I see, we have 3 stories in there. 20:23:17 MoC: todo++ fix the formatting of the user stories? :) 20:23:34 sure 20:23:48 cool. 20:23:56 do we want to discuss the usertag proposals now, then? 20:23:57 actually the sub lists are kind of integrated into the doc more or less 20:24:19 intrigeri: i'd say yes 20:24:31 ok, go. 20:24:35 first is about "When I want to ship an AppArmor profile in "my" package" 20:24:49 #action MoC fix user story formatting on https://wiki.debian.org/AppArmor/Progress#User_Stories 20:24:52 based on the stories i see 1 usertag: "apparmor" ;) 20:25:19 h01ger: i have understood that because we have the user pkg-apparmor team, adding apparmor as a tag is not useful 20:25:38 MoC: we need to set some tag 20:25:50 at least one 20:25:52 h01ger: what part of which user story are you talking of? 20:25:59 (cboltz: no PMs please. I will fix your orthographic issue, np) 20:26:18 intrigeri: ok, i see 2 tags now: profile-missing and profile-broken 20:26:30 * intrigeri confused. 20:26:46 no, there are 2 profiles 20:26:51 new-profile and ux 20:27:10 i propose to add "diagnose" - when you need help saignosing if a bug is indeed due to apparmor 20:27:26 and we talked about "merge-with-upstream" 20:27:35 based on the mysql profile issue 20:27:36 MoC: I wanted to avoid public "blaming" for minor issues in the middle of the meeting ;-) - but if you want everything in public, I'm fine with that 20:27:39 * intrigeri totally confused. 20:27:52 intrigeri: about what? 20:27:53 there are 3 stories and now i even/finally see 3 tags: aa-profile-missing, aa-profile-broken and aa-more-info 20:28:17 MoC: this discussion. the way draft tech solutions are mixed in the middle of user stories. 20:28:34 intrigeri: ok, let me fix the formatting first then? 20:28:40 h01ger: I don't think "aa-" is useful given these are *user*tags set for pkg-apparmor-team. 20:29:09 MoC: I think so. And please move everything that's not part of these stories (and not properly phrased) outside. Let's keep the solutions elsewhere. 20:29:10 intrigeri: probably you are not confused, but rather we are, and this confuses you. could you re-propose your proposal how to proceed now? 20:29:21 intrigeri: probably, yes. (aa-prefix) 20:29:51 * h01ger waits for the formatting to be done 20:30:04 ok done 20:30:40 still one glitch 20:30:42 thx 20:30:48 +thanks indeed 20:30:49 let me fix it better, sorry 20:30:58 "proposed usertag: merge-with-upstream" is still in the middle. 20:31:23 and some stories have "As a Debian package maintainer, ", some others not. 20:31:33 i was just editing this 20:31:38 :) 20:32:30 #save 20:32:30 so, now i basically put it the exact same way as proposed during the last meeting 20:32:47 I added the second one though 20:32:59 (except we've lost the
, but I'll stop now :) 20:33:31 now, I think we can take each of those and stuff the user stories a bit as needed, and maybe discuss the tech solutions if we have time and feel it can be quick? 20:33:39 #info the user stories should be put into super formalized language, cucumber style, eventually 20:33:59 anyway. that's a tool. let's now use it :) 20:34:01 intrigeri: discuss tech solutions or usertags? 20:34:06 * h01ger thinks this proper language/formatting really helps 20:34:11 MoC: usertags are one kind of tech solutions :) 20:34:38 user stories are meant (IMO) to take a step back, think of what "user" experience we want to offer, 20:34:38 wait a second. we are at topic "status updates" and i dont think we are really there anymore... 20:34:45 and from *there* think about solutions. 20:34:52 h01ger: true 20:35:00 h01ger: status update is done i guess? 20:35:08 i'm fine with switching topic if there is no more new status 20:35:13 yep. 20:35:19 #topic user stories 20:35:39 so, the 1st one is: 20:35:48 When I want to ship an AppArmor profile in "my" package 20:35:49 Then I need to be told how to do that 20:35:59 #info intrigeri> user stories are meant (IMO) to take a step back, think of what "user" experience we want to offer, and from *there* think about solutions. usertags are one kind of tech solutions :) 20:36:29 so, (as a pkg maintainer) I need to be aware that there are guidelines, doc, and a team to help me, basically, no? 20:36:37 intrigeri: yes 20:36:50 so it requires guidelines, doc, a team, and enough publicity about it. 20:37:00 * h01ger thinks we should discuss one story here in detail and let MoC work out the rest til next meeting? 20:37:02 MoC is working on the guidelines and doc (\o/) 20:37:37 there is a team, yes guidelines and doc are work in progress. publicity: planet? 20:37:39 h01ger: I'd like to discuss another story that possibly involves usertags too, as I'm in doubt we're on the same page regarding them. 20:37:49 intrigeri: ok, cool 20:37:58 * h01ger just wants this meeting to finish in 23min :) 20:38:00 MoC: d-d-announce, planet, Jessie release notes, etc. :) 20:38:18 h01ger: ah, OK. please tell us when beginning the meeting next time. 20:38:45 any other solution for the 1st user story? 20:38:57 DPN is also good for publicity 20:38:59 (we can brainstorm more in depth later, when it'll be time to do the publicity thing) 20:39:05 ok, intrigeri. do you think there should be an action item (on the long run) about publicity involving what you just said? 20:39:17 what I meant to arrive at is: no need for bug reports nor usertags, to solve this 1st user story :) 20:39:23 yep 20:39:24 and ack on defining meeting length at the beginning. i was blindly assuming "the default" ;) 20:39:25 ok 20:39:31 h01ger: :)) 20:39:46 MoC: yep. not necessarily as part of your project. I think you're doing everything so that we (the team) can publicize that once you're done :) 20:39:55 intrigeri: ok 20:39:58 hehe 20:40:00 so, next stepry. 20:40:02 story. 20:40:07 As a Debian package maintainer, 20:40:12 #save 20:40:33 (skipping the 2nd user story that's less complete than the others, for now) 20:40:39 ok 20:40:49 Who ships an AppArmor profile in "my" package, 20:40:53 When I receive a bug report that might be caused by AppArmor, 20:40:54 Then I want to 20:41:04 a) be told how to diagnose if AppArmor is involved 20:41:14 b) if yes, be told how to put the problem under the pkg-apparmor team's radar 20:41:27 a) is part of the documentation 20:41:40 b) requires usertagging 20:41:45 so, for (a), same as above (doc, and enough publicity so that people actually *know* that it might be caused by apparmor, and that there *is* doc) 20:41:53 MoC: agreed. 20:41:54 yep 20:41:57 #info check the full log for discussion on user stories and what follows up on them 20:42:24 are we good here? 20:42:41 not completely 20:42:51 ack. please go on then 20:42:51 so, we need two usertag here, right? 20:43:02 so, the usertag should simply involve adding pkg-apparmor-team@ as user 20:43:12 one for maintainers who don't want/can't debug themselves. one for once the AppArmor profile has been proven to be part of the problem. 20:43:26 and something which indicates that i need help with debugging 20:43:29 MoC: I think a usertag needs both a user and a tag. 20:43:37 he yes 20:43:47 no 20:43:51 maybe "help" and "buggy-profile"? 20:43:58 the user should always be the same. 20:44:03 ah, no, "help" is defined globally. 20:44:12 h01ger: I don't think anyone said the contrary :) 20:44:23 the user who "owns" all the aa-usertgs 20:44:35 intrigeri: ok, good, then i might have misunderstood. 20:44:40 h01ger: I was answering "the usertag should simply involve adding pkg-apparmor-team@ as user", which is wrong, as it needs a tag name (I think) 20:44:54 h01ger: sure, all our usertags will have pkg-a-t@alioth as the user. right? 20:45:07 yeah. whoever :) 20:45:08 intrigeri: and i was simply starting this sentence, add a user AND a tag, i wanted to say :)) 20:45:31 there's a global "help" tag on the BTS, right? so we can't use it as a usertag. 20:45:35 MoC: no, you add a usertag 20:45:38 * intrigeri verifying. 20:45:42 intrigeri: yes, there is help 20:45:46 OK. 20:46:00 maybe it also works as usertag 20:46:03 needs-help? 20:46:13 that works 20:46:14 h01ger: no, never reuse global tags as usertags. 20:46:20 point 20:46:26 h01ger: (the BTS is buggy if you do that.) 20:46:33 (totally buggy. useless.) 20:46:40 needs-help sounds good 20:46:49 "help-needed" is probably clearer and nicer. 20:46:59 * h01ger nods intrigeri 20:47:02 but then buggy-profile might not be necessary 20:47:06 cool. 20:47:23 "profile-kaputt" 20:47:39 MoC: I find it useful to be able to sort apart the two cases: we have a confirmed bug in our stuff, or we have something that *might* be caused by our stuff. 20:47:48 ok indeed 20:48:05 #i nfo anything? or next topic? or? 20:48:15 I don't like buggy-profile very much. There might be AppArmor issues caused by something else than a profile. 20:48:33 apparmor-broken? 20:48:44 well, it's more that it breaks something else. 20:48:53 let's leave this one for MoC to brainstorm and propose something? 20:48:58 ok 20:48:58 and next topic 20:49:16 #topic docs 20:49:18 #save 20:49:24 (trying to rush the meeting to its end while still trying to take useful things out of it.) 20:49:27 #agreed on usertag help-needed for package maintainers who cant debug their profile 20:49:54 #agreed we need another usertag for pkg maintainers who have identified that AppArmor is breaking something for them 20:50:05 #agreed U will brainstorm usertags (+stories) and propose something 20:50:33 if we switched # topic now, we would have even nicer formatting in the automated protocol, but meh :) 20:50:34 #action brainstorm on 2nd usertag for story nr. 3 once the AppArmor profile has been proven to be part of the problem 20:50:40 (re) 20:50:45 #topic docs 20:50:49 whats about this topic? 20:50:50 what's this exactly? 20:51:00 MoC proposed it 20:51:20 i think that was that the docs need some review 20:51:25 but maybe that is not urgent yet 20:51:41 intrigeri: you somehow said you could read the stuff over and report back to me? 20:51:47 I will, sure. 20:51:54 (and next time I'll do it before the meeting.) 20:52:00 * h01ger is happy to review if you send a dedicated mail "please review $this-doc" 20:52:07 ok 20:52:19 i need some pull request like that :) 20:52:32 #action intrigeri: review doc on debian wiki 20:52:48 MoC: ETA January 2 is good enough for you? 20:52:49 #action MoC send doc review requests to h01ger by email 20:52:51 ah, you mean AppArmor/PackageMaintainers ? 20:52:55 intrigeri: yes 20:53:03 i was about to propose this too 20:53:17 h01ger: yes, for example 20:53:29 #agreed intrigeri and h01ger review https://wiki.debian.org/AppArmor/PackageMaintainers til january 2nd 20:53:32 MoC: org-mode++'ed 20:53:42 MoC: what else? ;) 20:54:00 in docs, nothing 20:54:03 todo? 20:54:05 ok. 20:54:09 hmmm 20:54:13 [21:53] < MoC> | h01ger: yes, for example 20:54:19 so to review only that now? 20:54:58 I plan to review the 2 blog posts too. 20:55:05 intrigeri: that's great! 20:55:11 i like that 20:55:34 its basically our job to review everything :) 20:55:38 I'd like to see the status report shipped with proposed plans for the next 2 weeks. 20:55:45 intrigeri: i'd actually prefer if one of you reads them first before they appear on planet 20:55:45 #save 20:55:53 #topic todo 20:56:10 h01ger: ^ don't add to planet before one of us has reviewed. 20:57:06 intrigeri: no, MoC shouldnt push her posts to her blog before reviewing if she doesnt like. all new posts are automaticlly aggregated now. thats how planet works 20:57:25 h01ger: she published this *before* it was added to planet. 20:57:34 so, for the todo: i should continue working on the user stories and tags. documentation. fix the pidgin profile. pull pidgin-blinklight modification from upstream into apparmor-profiles-extra and work on the UDD query 20:58:04 h01ger: too late? OK. 20:58:11 intrigeri: i didnt put noone on planet without asking frist... 20:58:13 first 20:58:32 MoC: agreed :) 20:58:41 about the planet : please take it down if you want to, until you reviewed it. it might be better not to add strange things to this planet. 20:58:47 h01ger: sure. I thought you wouldn't do it during the meeting. sorry for the confusion. 20:58:59 its impossible to take down stuff from planet once its there... 20:59:05 however, i will publish the stuff because i need to do that every 2 weeks anyway 20:59:08 (and its fine on planet. i read it...) 20:59:27 what we can do is change the feed URL and i will add the posts to a specific category once they have been reviews 20:59:32 intrigeri: no done this afternoon already :) 20:59:34 reviewed, what do you think? 20:59:39 MoC: why? 20:59:46 MoC: makes sense. 21:00:12 if you want your post reviewed before posting, do that. once you post them on your blog, they are on planet. or do you want to introduce a preview category on your planet? 21:00:20 h01ger: so, it can appear on the planet of the outreach program first. and once reviewed it can appear on debian's planet 21:00:24 imo thats overkill, but fine. 21:00:39 MoC: but why? 21:00:43 so if we change the feed url to the reviewed category.. 21:00:55 nmistakes are fine 21:00:59 mistakes are needed 21:01:02 h01ger: because then we are sure that the communication about aa in debian is correct and nt faulty 21:01:07 how shall we learn without mistakes? 21:01:15 h01ger: it's easier to tell that from your position, than to experience it from MoC's. 21:01:19 h01ger: just saying. 21:01:47 "but fine" 21:01:49 h01ger: imagine i say something which is not correct, and then people who want their profile shipped have false info 21:01:56 that's the usecase i am tinking about 21:01:59 +h 21:02:15 then a bug will be found, reported and fixed. 21:02:44 MoC: if you want a different feed of your to be aggregated on planet, i'll happily change it to that 21:02:55 but wouldn't it be easier to have the right info right away? 21:03:19 h01ger: ok, let's do that, we can still change it, right? 21:03:39 i'll send you an email later with another feed url 21:03:43 * cboltz wonders if MoC's blog allows a public preview link (sent around via mail) before hitting the "publish" button 21:03:47 well, mostly you will post right stuff right away :) and if not, add a small note with a correction. or write a big bang post about it :) 21:03:49 MoC: sure 21:04:05 I see two remaining topics. 21:04:11 1. plans for next 2 weeks 21:04:13 2. next meeting 21:04:15 todo and ? 21:04:17 right 21:04:19 [21:57] < MoC> so, for the todo: i should continue working on the user stories and tags. documentation. fix the pidgin profile. pull pidgin-blinklight modification from upstream into apparmor-profiles-extra and work on the UDD query 21:04:28 do we agree on this todo? :) 21:04:36 it sounds good to me 21:05:03 not convinced by the UDD query, but that's probably because I'm lacking tech background and had no time to look at it / ask questions. 21:05:19 ACK the rest! 21:05:32 and "work on making usertag visibly" 21:05:34 MoC: once you have stuff ready for review in the wiki, please ping us. 21:05:40 intrigeri: ok 21:05:52 MoC: I see the notificaitons, but never know when it's time for me to look at it. 21:05:56 (many notifications :) 21:05:56 intrigeri: agreed on "...and work on making usertag visibly"? 21:05:58 intrigeri: and about UDD, if you manage eventually to read the progress page, you can still comment on it 21:06:06 h01ger: not sure what it means. 21:06:30 MoC: I've read it. 21:06:41 instead of working on the UDD query (a correct implementation proposal), she will work on... making usertags more visible. 21:06:47 MoC: let's discuss it later, e.g. after the meeting or before you start working on it. 21:06:47 though there are urls 21:06:50 #action MoC until january 2nd continue working on the user stories and tags. documentation. fix the pidgin profile. pull pidgin-blinklight modification from upstream into apparmor-profiles-extra 21:06:57 intrigeri: ack 21:07:02 hehe, thanks, agreed 21:07:19 #agreed we agreed on the last #action :) 21:07:22 #agreed on rediscussing working on the UDD query 21:07:27 #topic next meeting 21:07:55 can we also agree on trying to have 60min meetings each time? 21:07:57 #agreed MoC needs to ping intrigeri and h01ger explicitly once documentation is ready for review 21:07:59 I can do Jan 5-7 21:08:02 h01ger: +1 21:08:26 intrigeri: 1h good for you too? 21:08:28 7th i'll have an operation and will probably be sick for 2 days 21:08:30 happy to have this as a general goal. would rather schedule a bit more time just in case. 21:08:31 jn 5-7 works for me 21:08:41 so rather 5th if possible 21:08:53 ok, 5th be it. which time? 21:09:09 4pm? 21:09:13 ack 21:09:15 planning for 90min sounds doable for me, though i find long meetings tendious 21:09:39 I've found this meeting tedious, but not really because it was long. 21:09:48 i rather prefer to schedule a short ad-hoc session afterwards and keep meetings shorter+more focussed 21:09:52 so, Jan 5th, 4pm CET. 21:10:10 ack 21:10:13 #agreed next meeting , january 5th, 1600 CET, monday 21:10:16 cool. 21:10:21 :) 21:10:26 60 or 90min? 21:10:34 60min preferred 21:10:43 everybody prefers 60 minutes. 21:10:47 hehe 21:10:52 (of course.) 21:10:52 "we reserve 90m time, but plan for 60min meetings"? 21:10:53